XML-based Heterogeneous Data Integration On The Database Grid

Database Grid is a kind of Data Grid mainly for integrating database into Grid environment and providing effective publishment,access and administration for distributed heterogeneous databases resources.Now,the researches on database grid is at the underway and developing stage,the correlative implement technologies are immature, and there are no whole standards and specifications,so our research and practice work in this proiect is worth of exploring.The paper researches the heterogeneous data integration on the database grid.The integration work must resolve the problems such as complexity of grid environment, heterogeneity of data sources and global access,so as to provide users a global transparent query function.So the paper focuses on the researches about schema integration,query translation and data transformation.Schema integration is the foundation of data integration,and the paper discusses the hierarchy of schema integration,generation of partial schema & global schema and solutions of schema conflict;the data sources related to the user query must be found in the query decomposition,and the paper adopts the selection algorithm based on the model of predication correlation.The XQuery query should be translated into corresponding SQL query correctly and efficiently.The paper gives a solution which has binding variable tree as a middle result to solve the problem. And the algorithm BVT is used to create the binding variable tree,while the algorithm GetSQL integrates the scattered SQL into the final SQL query.The results should be returned in XML format,so the relational data must be transformed into XML data.The paper completes the transformation by the rule of RelationalXML.The paper implements a book union-query system basing the multi-libraries which completes the data integration.The system provider enough metadata induing user metadata,data metadata,deployment metadata;Global data mode is created then, supporting global data access;The global query for the global data mode is then decomposed,transformed and executed,and the query result is transformed into the XML format,and the integrated result is returned to user lastly.So the data integration is completed,and the system can applicable for practical work.